YOLOv10, model.pt –> model.onnx, Dynamic Batch Size + NMS

[Egorundel]

Hello, do you know how to use PyTorch weights (model.pt) to make ONNX with dynamic batch size and NMS module for YOLOv10?

I tried this repository:
https://github.com/laugh12321/yolov10/tree/nms.

And it seems to work, but I have a feeling that there is some kind of problem, so why can’t I do what I described in this issue: NVIDIA/TensorRT#4074